Here is my five cents of worth.
Just to make everything clear. You use the same system set-up, and load 1,5 mL sample using a 5 mL loop and where the only difference is the column dimension and the flowrate. If so, you have several effects playing with you. Now, if I've done my calculations in a correct manner.
1. on the 100x10, you load a fairly large sample volume (approx. 19 vol-% of the total column volume) using a sample loop that has almost the same volume as the void volume of the semi-prep column. At your reasonable flowrate it takes 1.67 min to elute the sample from your loop, giving a t'0 of 3.2 minutes. Your analyte elutes at 12 minutes, thus with a retention factor of 2,7.
2. Then you switch to a 100x4.6 mm column, where the total column volume is 1.66 mL and the void volume is 1.0 mL. You load 1.5 mL approx. 90 vol-% of the total column volume using a sample loop that has 5 times larger volume compared to the void volume of the analytical column. Now, as the flowrate you employ is scaled from the 10 mm id, your analyte (with a retention factor of 2,7) should elute around 37 minutes. In my mind, the extra-column effects in this second setup should be so severe that I can't really image that you would be able to get any results at all. 3. Only the band-broadening effects inside the sample loop should be so severe that it would prohibit data retrieval. It takes more than 8 minutes to empty the loop and I guess the diffusion must be rather substantial.
You must scale down the amount and volume loaded onto the analytical column to minimise the extra-column band-broadening effects and prevent volume overload too.
My personal rule of thumb which is also my reccomendation to other chromatographers:
Inject upto 1% of the column volume with full chromatographic efficency.
Inject 10 % of the column volume and achieve only 20 % of the full chromatographic efficency.
I have some slides available if it is of interest, otherwise it can be found in our HILIC guide.
